In microwave communications systems and components, a linear phase response, i.e. so called consistent group delay, and a flat amplitude response are necessary in order to keep the waveform of signals transmitted through or by the system or component. Without consistent group delay and flat amplitude response, distortion will occur in the waveform, resulting in inter-symbol interference.
For example, in ultra-wideband, UWB, applications, consistent group delay is a highly important issue, since UWB systems use impulse signals to transmit information, and variations in group delay may cause fatal errors.
Another area of technology where group delay is also highly important is wideband power amplifier design.
Variations in group delay mainly originate from amplifiers, filters and load miss-matching in the system or component. In a typical low-pass or band-pass filter, it can be found that the main group delay variations occur near the edge of the pass band, i.e. in “transition areas” of the filter. Such group delay variations need to be compensated for. In order to accomplish this compensation, different technologies have been used: in analog systems, especially at microwave frequencies, all-pass networks are usually used to generate positive group delay signals, i.e. the opposite of negative group delay.
As an alternative to IIR and FIR filters, negative group delay (NGD) circuits can be used in order to suppress the group delay which increases near the edge of the pass-band of a low-pass or band-pass filter, without changing group delay in other ranges.
If all-pass networks are used in order to equalize group delay, passive components with a high quality factor, Q, are necessary, since otherwise a large attenuation would occur at resonance frequencies. Typically, the Q factor for “lumped components” in such solutions should be larger than 300, which is a very high value, particularly in monolithic microwave integrated circuit, MMIC, technology. Also, a single all-pass network has a limited frequency bandwidth and group delay variation, so in order to cover an entire pass-band, several cascaded all-pass networks with different center frequencies would be needed.
If a negative group delay circuit is used for equalizing group delay, a single stage circuit is usually enough to compensate for the group delay increases which occur at the edge of the pass-band. Unfortunately, a large attenuation in magnitude cannot be avoided, which has to be compensated for, usually by means of amplifiers. However, the gain increase obtained by using amplifiers is not limited to the frequency range where the negative group delay occurs, which is obviously undesired.